OKUJI Yoshinori wrote: > Another issue: I read VBE/Core Standard 3.0 and thought that the > proposed information to be passed to a Multiboot kernel wouldn't be > enough. ... snip ... > This structure doesn't have information about the color format (direct > color, packing color, YUV color, etc.), the logical scanline length, > or the mask size of each of red, green, blue. Thus, I feel that GRUB > should pass the mode information returned by VBE function 01h. What do > you think? Looking at VBE3.0 specifications, I would suggest that following information would be given: PMInfoBlock (page 21) (if found) ModeInfoBlock (page 30) (current video mode information) PMInfoBlock would give pointer to protected mode entry point which is handy for protected mode OSes. And also couple of handy selectors. ModeInfoBlock would allow OSes to fully continue work on current video mode, eg. if was in video mode then it could continue with graphical interface which would give much nicer look at startup. And combining these two there is no other need to make OS to call real mode BIOS calls.
